Evaluating deep learning architectures for Speech Emotion Recognition

By Haytham M. Fayek, Margaret Lech, Lawrence J. Cavedon. Neural Networks, 2017

540 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Experimental and Cognitive Psychology, Signal Processing

View Paper: https://www.semanticscholar.org/paper/caffe624694d7b015d5f4be983e41247d04eee38

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Experimental and Cognitive Psychology

In Signal Processing

Citations in the corpus

Short-term water quality variable prediction using a hybrid CNN–LSTM deep learning model (2020, background): “2019), speech recognition (Fayek et al. 2017; Cummins et al. 2018), time series and spatiotemporal prediction (Chen et al.”
Emotion Recognition from Speech Using Wav2vec 2.0 Embeddings (2021, methodology): “In this work, we used 4 emotional classes: anger, happiness, sadness and neutral, and following the work in [36], we relabeled excitement samples as happiness.”
A review on the long short-term memory model (2020, background): “In speech data emotion recognition, Fayek et al. (2017) conducted a review study to compare various deep learning architectures, both feed-forward and recurrent networks.”
Introducing AmuS: The Amused Speech Database (2017, background): “The latest works on recognition using audio cues rely on feature learning systems using the power of deep learning to learn an internal representation of raw data with respect to a given task [17].”
Reinforcement Learning for Primary care e Appointment Scheduling (2017, methodology): “Chapter 3 – Reinforcement Learning 18 recurrent architectures, sigmoid functions, such as the logistic function, s(x), presented in the next equation, as well as the hyperbolic tangent (TanH) function, are frequently used as activation unit functions, instead of ReLU, since the latter amplifies the ”
A Systematic Review on Affective Computing: Emotion Models, Databases, and Recent Advances (2022, background): “However DLbased SER constructs an end-to-end CNN architecture to predict the final emotion without considering feature engineering and selection [164].”
A CNN-Assisted Enhanced Audio Signal Processing for Speech Emotion Recognition (2019, methodology): “Haytham et al. [38] described a neural network and RNN-based technique for SER using spectrograms of frames to train DNN, which has a high computational coast and did not achieve high accuracy.”